Exwayz

Take a guess on where this Point Cloud video was taken 👇 (Tip: It was taken while visiting Visimind AB 👀) Special thanks to our partner Exwayz for capturing this video with our Falcon #LiDAR!

I am delighted to show you an example of Exwayz SLAM CLI usage! 🚀 exwayz_slam is the first executable of our #3d mapping software suite 👨💻 Its goal is to process raw LiDAR data captured in native #PCAP format or #ROS bag format: it computes what we call the LiDAR #odometry, which is the relative trajectory of the sensor from its first position in the dataset. On the video below, this trajectory is displayed in white. Thanks to this estimated trajectory, a point cloud #map is simulatenously built (in viridis colormap below) by aggregating the LiDAR frames on the corresponding trajectory points. Basically, a good SLAM produces a locally sharp point cloud, as it is the case with ours of course 😃 The video below is not accelerated: on my computer, a laptop with an Intel core i7-12700U, exwayz_slam processes data at ~120Hz whereas the acquisition frame rate was 20Hz, which means that the processing is 6 times faster than the data collection ⏩ Usefull when you have several ours of data to process! The other Exwayz executables from Exwayz #3DM allow to compensate for the drift with our loop closure solution, and to do fine georeferencing. I'll dedicate a post to them one day, stay tuned 🔥

Two months back, when we were in Vegas for the #CES2024, we deployed our own #dual sensor mapping setup. As shown in the photo below, we put one sensor horizontally, and the second one with a strong 60° tilt. One sensor alone already measures high quality and abundant data, thus why using an additionnal one? 🤔 Actually, the sensor mounted horizontally covers decently the surroundings of the vehicle, which is a good property for doing LiDAR #slam, while the tilted sensor allows to measure points that are high on the closeby structures, and directly adjacent to the vehicle. One could think that using the tilted sensor alone would be sufficient, as it gathers the most valuable data for mapping purposes, however, the #LiDAR SLAM can get lost with a very tilted configuration because of the lack of features measured by the sensor (even Exwayz cannot do magic when there is not enough geometry 🔮). We already discussed that in an older post : https://lnkd.in/gZvC6trP Doing sensor fusion is the key to get the best of the two worlds 💡 The idea is to fuse the two point cloud feeds to create "virtual frames" composed of both LiDAR measures and to use this as the input data for the #SLAM. However, doing this LiDAR #fusion requires 1) excellent time synchronization 2) phase locking the sensor 3) super accurate extrinsic calibration. The two first points are mainly hardware challenges (thanks Hassan 🤖 Bouchiba and Antoine Plat for dealing with that 😅) that require playing around with the sensors setup and network configuration. Having the time synchronization and phase locking is necessary to know how to compose the frames of each sensor, based on the time of each point. Regarding the sensor extrinsic calibration, it can be delt with in multiple ways, but it is worthy noticing that a tiny 0.5° error in the #calibration between the two sensors can already lead to 40cm offsets between points measured at 50m. That’s why being able to compute this #calibration super accurately is critical for the overall quality of the point cloud. Fortunately, Exwayz SLAM technologies and clever ideas allow to compute LiDAR - LiDAR calibration in a software fashion that is able to reach such lever of precision 📐 Once these challenges are overcome, we can create centimer-accurate, exhaustive, and reliable 3D point clouds with a processing time that is 5x shorter than the acquisition duration! Ready to map cities! 🏙️ #LiDAR #SLAM #dual #sensor #setup Aubin Tchoi

À Las Vegas, la startup française Exwayz a proposé à Numerama de monter dans une voiture équipée de deux capteurs LiDAR, afin de nous montrer comment elle

Final Day of #CES ! Great to see so many familiar faces. It was a perfect time to consolidate our partnerships with key 🔑 actors of the ecosystem. We were able to showcase our latest innovations to the world 🌎. Our tech keeps getting better and better 🚀 ; for this edition of #CES we conducted some live demonstrations with a dual-sensor setup, that some of you were lucky enough to check out. More details on it in a later post!

There is definitely a lot happening in my first year at Exwayz, and I am currently at #CES 2024, attending as an exposant to exhibit our tech to the whole wide world 🌍 This is not only my first time being at Vegas ✨️, as this is also my first tech event, and it seems that we are starting off on the main stage! 🎉 After one day and a half of show, I can say for sure that this truly feels like the place to be for someone who works in robotics and #LiDAR processing. It’s incredible to see how many LiDAR sensors fit in one place, and it’s great to meet the people behind some truly amazing innovations! 🤖 Come meet us on booth 4025! Hassan 🤖 Bouchiba Mathias Corsia Antoine Plat

🏁 Can our SLAM follow an #IndyCar running at > 140 mph? 🤔 We are always on the look for new challenges here at Exwayz! Today I want to share with you a follow up of a post I published a while back on the #RACECAR dataset 🏎️ (link in the first comment 👇). In this previous post, I showed the 3D maps we were able to obtain on the legendary racing circuits of Las Vegas Motor Speedway and Indianapolis Motor Speedway 🏁. Today I will show you how our SLAM algorithm can localize a vehicle running at speeds exceeding 140 mph (225 km/h) 🥵 within these maps! Check out this preview that showcases our method in action by replaying the data and localizing the running IndyCar in real time ⏱️. You can try to imagine what it would feel like to find yourself in the cockpit thanks to our awesome widgets 😎, especially the ones with the speed and the acceleration: 2G is a strong kickback! #SLAM methods typically struggle with acquisitions performed at such high speeds. The most obvious source of difficulty lies in the large displacement observed between two successive frames as it lowers the overlap between them. Another major phenomenon is the so-called ego-motion distorsion: each single frame is internally distorted due to the motion of the vehicle during their acquisition. In the displayed data, more than 10 feet (3 meters) separate the positions of the LiDAR sensor at the beginning and end of each frame. This inner warp has to be accounted for in a joint manner with the registration: indeed, the odometry is required to performed the ego-motion compensation, which in turn benefits the registration by providing more realistic data. Looking forward to seeing what the Indy Autonomous Challenge has cooked up for #ces2024, which we will attend with the whole team! See you there, and stay tuned for more tech content! 🌟 Hassan 🤖 Bouchiba Mathias Corsia Antoine Plat

Meet us at #CES 2024 in Las Vegas! 🌟 We will show our very latest innovations with live demos at our #booth 4025 on the Automotive #french pavillion. From city scale #mapping to real-time #navigation in challenging scenarios, come to discover how our technology will revolutionnize autonomous system deployment! 🤖 Antoine Plat, Hassan 🤖 Bouchiba, Aubin Tchoi and Mathias Corsia will be pleased to receive you 😉 Thanks again to #BusinessFrance who invited us, we can't wait to be there!

The Forum #MVA took place this morning at the ENS Paris-Saclay 🧑🎓. With Mathias, we had the opportunity to present our internships at Exwayz to the students of the master, which is known worldwide for its excellence in the fields of Applied Mathematics and Computer Vision. It was a long ride 🚌 but was definitely worth it! 😄 This event had a very special meaning for me, as I attended this forum a year earlier as a student of the MVA, and met with the team at Exwayz for the first time! Coming from a background in optimization, I discovered the world of 3D Computer Vision 🌐 for the first time back then and look back dearly on this time. It was an enriching experience to find myself on the other side of the table this time and I was particularly pleased by the enthusiasm of all the students we had the opportunity to chat with over the course of the morning 😄. As a math lover, I am more than happy to watch this master grow each year and consistently provide high-quality math courses that help build up strong theoretical foundations for students who learn to think straight and solve complex problems with a scientific approach. Such an inspiring place to be! 👏 Hassan 🤖 Bouchiba Antoine Plat

I’m thrilled to announce that Exwayz has been selected in the prestigious #Future40 #startup ranking of STATION F 🤩 This really means a lot to me and to our team. When I founded Exwayz with Mathias Corsia and Antoine Plat, it was not easy to find a relevant incubator for our robotics project, where we could be advised by entrepreneurship experts and be part of a dynamic ecosystem. In January 2022, integrating the #HEC program of STATION F was already a big achievement and really boosted our development, but at the time we were far from imagining that we would be among the 40 best start-ups of the campus one day 🚀 This makes us realize how far we have come, but we don’t forget that the journey is still long and full of exciting adventures!! 🤖 Let’s shape the #Future40 of mobile #robotics together and big thanks to to Roxanne Varza Marwan Elfitesse Loup Audouy, all the STATION F team and all the people who support us and make this project so cool 😁 Aubin Tchoi #slam #lidar